The increasing interconnectedness of the world's economic, political, and social systems have made individuals' perceptions of other national and cultural groups increasingly consequential. More and more of the decisions that individuals and societies make have international ramifications. Although international travel is increasing, many individuals' knowledge of people of other societies is heavily dependent on the mass media. For many, large portions of the world are represented by what is seen on TV and film screens and what is read in the newspapers. Given the mass media's function as a means through which some audiences' come to know what people of specific societies are like, the content of media representations is vitally important. The patterns that appear in the media can help determine the nature of the audiences' beliefs about particular societies.
The premise that media representations structure the perceptions of the audience is fundamental to the work of many scholars who have investigated how peoples from societies in South America, Asia, and Africa are represented in the media of North America and Europe. Media representations are seen as a link in a circular chain. They both reflect and reinforce existing patterns of power. One of the seminal works in this area is Edward Said's 1978 book, Orientalism. Said argues that Western scholars' understandings of the Middle East and India are constructed in ways that support the West's vision of itself and justify Western control of these areas. Other scholars have expanded on Said's work by further investigating mass media portrayals of the Middle East (Nadel, 1997) and by applying similar models to Western portrayals of the cultures of Africa, South and Central America, and East Asia (Torgovnick, 1990; Springer, 1991; Shome, 1996; Heung, 1997). Shohat and Stam (1994), for example, argue that Western media tend to portray people of Africa, Asia, and Central and South America as childlike, instinctual, and close to nature. The suggestion of the irresponsibility of individuals from these regions can be seen to justify the West's attempts to exercise control over these areas of the world. Furthermore, these kinds of portrayals allow for the role of the White, male European or North American as conqueror or sense-maker. The heroic European or North American who saves or rules over peoples of other regions is a common figure in films and adventure novels created in the West but set outside it.
Experimental researchers have found evidence that, at least in certain circumstances, media representations can shape international audiences' perceptions of the societies from which the media come. Most of this work deals with the impact of exported US television programming on perceptions of the United States. Several studies have found exposure to US shows to be associated with perceptions of the United States that concord with the content of American programming. Tan, Li, and Simpson (1986), for example, found that viewership of "Dallas" and "Dynasty" was associated with perceptions of people of the US as materialistic among Taiwanese viewers. In Mexico, viewership of "Dynasty" was related to perceptions of people of the US as individualistic and pleasure-loving, whereas viewing "Dallas was associated with understandings of people of the US as aggressive and cruel. Tan and Suarchavarat (1988), found a similar pattern of results in Thailand, as did Pingree and Hawkins (1981) in Australia, Weimann (1984) in Israel, Willnat, He, and Hao (1996) in Hong Kong, China, and Singapore, Saito (1996) in Japan, and Citipitioglu and Elasmar (1996) in Turkey.
In addition to the society-level impacts of media content patterns that have been explicated by cultural studies scholars, individual viewers' media-influenced perceptions of societies can also shape how members of these societies are treated when they are met face to face. Beliefs about the character of social groups can become cognitive schema that shape what is noticed about newly-encountered members of these groups (Rothbart, Evans, and Fulero, 1979; Cohen, 1981; Gaertner and McLaughlin, 1983; O'Sullivan and Durso, 1984; Devine, 1994; Gilbert and Hixon, 1991; Dovidio, Evans, and Tyler, 1986; Ford, Stangor, and Duan, 1994) and help determine how their actions or behavior are interpreted (Duncan, 1976; Sagar and Schofield, 1980; Darley and Gross, 1983).
However, the processes though which the media shape audiences' perceptions of the societies' represented in the texts are not absolute or uniform. As established by researchers such as Morley (1980), Livingstone (1990), and Liebes and Katz (1990), audiences vary in their understandings and evaluations of the same media text. This study seeks to explore the processes and factors that shape the way the position of the audience and the content of a text interact to shape the way media representations can impact audiences' perceptions of represented societies. It seeks to clarify the mechanisms through which audiences apply what they see in media representations to their understandings of the societies the media represent.
Previous work in intergroup perception has identified several cognitive heuristics that may influence the way audiences learn about societies from the media.
Character representativeness. One of these heuristics is that a group exemplar that is seen as typical of its group is more likely to shape an observer's perceptions of the group as a whole than an exemplar that is seen as atypical. This is perhaps illustrated most clearly by the work on counter-stereotypes. For example, Weber and Crocker (1983) found that a very atypical example had less of an effect on a stereotype's central tendency than several moderately representative ones. Similarly, Johnston and Hewstone (1992) found that counterexamples' influences on perceptions of their groups were predicted by their perceived typicality. In one of the most direct examinations of the effect of the typicality of counterexamples, Wilder, Simon, and Faith (1996) found that when counter-stereotypical examples were like other group members on all but a few attributes, perceptions of the group as a whole were less stereotypical than when counterexamples diverged on many attributes. Kunda and Oleson (1995; 1997) found similar results.
If this pattern holds in regards to media exemplars, it would suggest that audiences' perceptions of the representativeness of the characters help determine the media text's impact. The first goal of this study is to determine whether the attributes of representative media characters are more likely to be applied to viewers' real world understandings than attributes of unrepresentative characters.
Society Variability. A factor that can affect perceptions of a character's representativeness, in turn, is an audience's perception of the variability of the society portrayed in a media text. Variability perceptions are consequential to investigations of media learning because they make individual representatives of the group seem more informative about other members. This suggests that variability perceptions could modify the extent to which a media portrayal's attributes are extrapolated onto its group as a whole. A single example is more likely to inform perceptions of the rest of the group when group members are felt to be relatively invariable. A series of experimental studies supports the notion that the perceived variability of a group shapes the extent to which a single example affects observers' understandings of the group as a whole. Rehder and Hastie (1996) found that observers were less willing to adjust their estimates of the central tendency of a group in response to disconfirming examples when a group was seen as relatively diffuse than when it was seen as homogenous. Park and Hastie (1987) found observers were more likely to generalize from a single example to an entire group if the group was seen as homogenous. Single examples have less of an impact on outsiders' perceptions of the group when the group is seen as highly variable.
Category Structures. This work suggests that the nature of the society with which media characters are associated can impact a text's power to shape audiences' perception of the nature or attributes of that society. However, media characters, like people, are multi-dimensional. Even a brief portrayal offers a myriad of abstract concepts to which the attributes of a specific text may be applied.
Work on category activation in interpersonal environments suggests that priming a specific category structure leads to the greater impact of category understandings on individual members (Hilton and von Hipple, 1996). For example, Beckett and Park (1995) found that gender stereotypes had greater effects on evaluations of individual targets when observers were presented with photos that made the targets' genders more salient. Increased salience of a particular category structure seems to increase the power a single example has on observers' perceptions of the example's category as a whole.
The second goal of this study is to explore whether the activation of a particular category structure affects a media representation's power to shape audiences' evaluations of social groups in a similar way. The activation of a category structure may influence a text's impact on viewer perceptions of the real-world groups defined by that category structure. If a film is viewed in a context that stresses the society of the characters, it may be more likely to impact the viewers' perceptions of society than if another aspect of the characters' identities, such as their gender, social class, or sexual orientation, were highlighted. The activation of particular category structures is likely to affect the perceived variability of the groups defined by those structures, and therefore the impact of specific examples on group understandings. Therefore, the activation of a particular category structure should be associated with perceptions of the representativeness of the characters in reference to those categories. Category structure activation should also be associated with the impact of the characters on real-world perceptions of the group.
Category structures can be activated by the content of the text and by the environment in which it is seen. The importance of the various attributes that define the characters' group memberships is shaped by the nature of the narrative in which they appear and the way the story is presented. For example, the cast of the TV hospital drama "ER" can be sorted in terms of their profession, gender, race, attitude towards HMOs, or a dozen other sets of attributes. The relative salience of these category structures will change across and within episodes. When the nursing staff is on strike, the professional roles of the characters are likely to be very salient. Whether characters are administrators, doctors, or nurses will be closely tied to their attitudes and behavior in these circumstances. When the issue is whether a male doctor acted insensitively when breaking off his romance with a female hospital administrator, gender may become more salient. It may be more strongly correlated with attitude and behavior.
The media materials that surround a specific film or television text may also shape the salience of a particular category structure. The media are not consumed in isolation. Audience members often come to a text after having seen promotional materials or read critical reviews. Although little experimental work has been done on the way these materials shape viewer responses, it is likely that the way a text is presented can shape the way audiences respond to it. Categories of social membership, for example, are more likely to be salient to viewers when a film is presented as being about Taiwan than when the same film is presented as being about family relationships.
An important aspect of the environment in which a film or television program is seen is the promotional and critical materials that surround it. Most of the work on media promotional materials has been done in reference to theatrical films. Survey and interview research indicate that film audiences are heavily exposed to trailers, reviews, and advertising and that these messages influence their decisions to see particular movies (Custen, 1980; Austin, 1989). It is possible that these materials also shape the way a visual media text and the characters within it are conceptualized by cueing or priming specific aspects of the characters' identities.
I investigate these issues through an experimental study. The design of the study is a 2 X 2 manipulation. Half of the participants saw a segment of a film from a society of which they were a member, and half saw a film from a society with which they had no real-world experience. Before seeing the film, half of each of these groups read a summary of the film that was designed to activate category structures of nationality or society membership. After seeing the film segment, the participants were asked to report their evaluations of the film characters, as well as their perceptions of the societies from which the film came. I predicted that, for each film, the representativeness of the characters would predict the degree to which the characters' attributes match those of the characters' source society. The priming summary was designed to activate the social memberships of nationality and society membership. Therefore, those exposed to the prime should see the film society as less variable, and the films' characters as more homogenous than should those who were exposed to the control materials. I also predicted that the representativeness of the characters would be associated with the perceived variability of the society.
Sample One hundred and fifty-seven people took part in the study. The eligibility criteria were having been born in the US, US citizenship, and having learned English as a first language. Eligibility was defined in these terms because it was felt that the sense of membership and familiarity with other countries that is likely to be associated with having international familial or cultural ties could serve as a form of cross-categorization that would complicate the study. People were recruited to participate in the study from several colleges and universities near a large, northeastern city through classes, news-group postings, and newspaper advertisements.
The data of 18 people who took part in the study as part of a class exercise were excluded because they did not meet the citizenship or language eligibility requirements. The data of 19 participants were excluded because they completed the questionnaire incorrectly, leaving a final sample size of 120. Fifty-nine percent of the sample was women. The average age was 22.18(SD=5.17), with a range from 18 to 44. Eighty-three percent of sample identified themselves as White, 10% identified themselves as African-American, and 5% identified themselves as Asian or Asian American. Five participants, or 4%, identified themselves as Hispanic or Latino. Thirty-eight percent felt they spoke two languages "well enough to carry out a routine conversation." Four participants identified themselves as tri-lingual.
This sample had a relatively high exposure rate to movies. Seventy-two percent saw a movie, either in the theaters, on TV or on home video, at least once every 2 weeks. Ninety-two percent saw one a month. However, these participants, like US moviegoers in general, had much less frequent exposure to non-US films. Nine percent of the participants saw a non-US film once a month, and 51% of the participants estimated they saw at least two a year. Many of these films are likely to have come from Great Britain or Australia, as is indicated by the relative rarity of viewing subtitled movies. Six percent saw a subtitled or dubbed film at least once a month, 28% estimated they saw two a year.
In order to increase generalizability, two different film segments from two different societies were used in this study. One was from US film directed by Michael Steinberg called "Bodies, Rest and Motion." The second was from a Taiwanese film directed by Ang Lee titled "Eat, Drink, Man, Woman." The film segments were randomly assigned to screening sessions.
The size of the sample that saw the US film is 59. Five people had seen the film previously. Seven felt they were able to recognize the title or felt that they had heard something about it. On a scale from one to seven, the mean of the participants' agreement with the statement that they were "familiar with the culture in which the film was set" is 3.69(SD=2.12). The mean of their agreement with the statement that the "film was made with people like me in mind as an audience" is 3.17(SD=1.90).
Seventy-two participants saw the Taiwanese film "Eat, Drink, Man, Woman." Only three people had seen all or part of the film. Six felt they were able to recognize the title or that they had heard something about it. On a scale from one to seven, the mean of the participants' agreement with the statement that they were "familiar with the culture in which the film was set" is 2.78(SD=1.80). The mean of their agreement with the statement that the "film was made with people like me in mind as an audience" is 3.77(SD=1.70).
Procedure At the beginning of each session, all the participants were given a sealed questionnaire booklet and a consent form. They were asked to read, sign and pass up the consent forms. The top sheet of the questionnaire booklet contained a set of instructions. The participants were told that they would be watching the opening ten minutes of a mainstream, commercially released motion picture. They were informed that they would only be seeing a short clip of the film. Therefore, the first page of the questionnaire contained a brief summary of the film, like something they would find in a movie guide section of a newspaper. They were asked to read this summary because it would help them complete the questionnaire later. These instructions were repeated orally in all sessions.
Although the instructions were always the same, there were two different priming summaries used for each film. Each was formatted to look like a thumb-nail film review from a newspaper entertainment guide. One version stressed the national origin of the film, whereas the second version emphasized the thematic content of the film (see appendix). The questionnaires were randomly distributed so that half the participants in each session read each version of the questionnaire. When all the participants were ready to move on, they were informed that the clip they would be seeing was from the very beginning of the film, just after the opening credits. They were told where the film was from, the year it came out and, in the case of the Taiwanese film, warned about the subtitles.
After the screening, the participants were asked to open and complete the questionnaire booklet. The booklet began with measures that allowed the participants to rate the representativeness of two characters with the most screen time in the film segment. In the case of "Bodies, Rest, and Motion," they were asked to evaluate one male and one female character, called Nick and Beth, respectively. The participants were asked to estimate the characters' age. Beth was estimated to be 25(SD= 2.41) years old, whereas Nick age was estimated to be 27(SD=2.92). Those who saw the Taiwanese film, "Eat, Drink, Man, Woman," were asked to evaluate two female characters. One, Jia-Ning, was estimated to be 18(SD=2.15) years old. The other, Jia-Chien, was estimated to be 26(SD=3.32) years old.
Participants were then asked to complete a series of questions designed to measure the variability of the society in which the movie was set. Those who watched the Taiwanese film completed measures of their familiarity with Taiwan, whereas those who saw the US film were asked to evaluate the United States. Then both samples were asked "when they first became aware of where the film was made" as a manipulation check. The questionnaire concluded with items measuring their exposure to different kinds of movies and basic demographic questions. After they completed the questionnaire, all participants were debriefed. Participants who were not receiving extra credit were paid and asked to fill out receipts.
Character representativeness. Character representativeness is the degree to which the attributes of character are felt to be shared by others within the social group with which the character is identified. Participants were asked to evaluate the two characters with the most screen time. The order in which the characters were evaluated was randomized across participants. They were asked to think of men or women, as appropriate, of a particular character's "age and culture." Participants then asked to indicate on a seven-point scale "how typical or atypical [the character] is of this group." There was a marked neutral point at four. The participants were also asked to indicate, on a seven-point scale, their agreement or disagreement with the statement that the character was "very different from other men/women of his/her age and culture" and that "Many men/women of [the character's] age and culture would behave like he/she did if they were in his/her situation." These questions were embedded in a series of items asking participants to evaluate the likableness, realism, and humorousness of the characters and the degree to which they identified with them. These questions served as controls and helped mask the intent of the study. The representativeness items were re-coded for direction and averaged to create a single index. The reliability of the scale, including both characters in the film segments, was .739 for the US film and .623 for the Taiwanese film.
Application of character attributes to society perceptions. When the participants evaluated the films' societies, they completed a series of seven-point scales describing the "average" person of the society. The scales were marked with an attribute at one end and the opposite of the attribute at the other. There was a labeled, neutral mid-point. An attempt was made to include attributes found in previous research (e.g., Kippax and Brigden, 1977; Lalonde and Gardner, 1989; Haslam, Oakes, Turner, and McGarty, 1995) to characterize the four societies in the study as well as attributes that had been used by pre-test participants to describe the film characters. The attributes used in these measures were, polite/impolite, traditional/modern, unambitious/ambitious, impulsive/self-controlled, confident/unconfident, practical/idealistic, talkative/quiet, group-oriented/individualistic, happy-go-lucky/serious, shy/friendly. The order of the attributes was randomly determined, but was the same for all participants.
After evaluating the overall "average person," participants were asked to evaluate the average person within four subgroups for each society. The subgroups were young men (age 17-34), young women, mature men (age 35-68), and mature women. The order of the subgroups was randomly determined, but was the same for all participants.
Perceived variability of the films' source societies. Perceived variability was evaluated through three measures. The first measure, the range measure, was perhaps the simplest to administer and to calculate. Participants were presented with four unnumbered scales marked with attributes found by previous cross-cultural research to vary systematically across societies--self-sufficiency, reservedness, caution, and competitiveness. They were asked to indicate "where the average person in [the place] would be on each of the following attributes with an X." Then, they were to "indicate with two slashes (/) where the two most extreme people in [the place] would be, that is, the person who shows the characteristic most strongly and the person who shows the characteristic least strongly." An example was provided.
The perceived range of each attribute was determined by measuring the distance, in millimeters, between the marks representing lowest and highest person. The scale could vary from .1, for those who placed the maximum and minimum of the population's distribution .1 millimeter apart, to 15.2, for those who marked the maximum and minimum at the extreme ends of the scale. Range scores were averaged across attributes to arrive at a single measure for each evaluated society. This measure has been used successfully in a variety of studies including Park and Judd (1990), Judd, Park, Ryan, Bauer, and Kraus, (1995), and Ryan, Judd and Park (1996). The reliability across attributes for the range measure was .659 for those evaluating the United States and .806 for those evaluating Taiwan.
Perceptions of the variability within the source societies were also evaluated by two measures that were constructed from a single set of four scales, which were developed by Park and Judd (1990) from a measure created by Linville, Salovey, and Fischer (1986). Each of the scales asked participants to evaluate the target society on a different attribute. The attributes used in the scales were selected because they had been established by previous cross-cultural research distinguish cultural groups. These attributes are adventurous/unadventurous, competitive/collaborative, independent/dependent, and outgoing/reserved. Participants were provided with stick-on dots in three sizes and presented with four six-value scales. The values of the scales were labeled to indicate a range of degrees to which an individual might posses the attribute. For example, in reference to the "adventurousness scale," one end was labeled "adventurous." The next values were labeled "moderately adventurous," "slightly adventurous," "slightly unadventurous," and "moderately unadventurous." The other end was labeled "very unadventurous."
Participants were presented with the scales and a supply of circular stickers. They were asked to think of a particular group as a whole. They were asked to "estimate the proportion of people from [the place] who fall into each level of the characteristics listed below". If they thought a lot of people within the group occupied a particular position on the scale, they were asked to mark it with a large sticker. If they thought a moderate number of people occupied a particular position on the scale, they were to mark it with a medium sticker. If they thought a small number of people occupied a particular position on the scale, they were to mark it with a small sticker. If they thought no one, or almost no one, occupied the position, they were asked to leave the value blank. An example was provided.
From these scales two measures, the probability of differentiation and perceived variability, were calculated. Probability of differentiation is the probability "that two randomly chosen group members will be perceived to differ in terms of the attribute" (Linville, et al., 1986, p. 167). Perceived variability is the extent to which category members are viewed as widely dispersed about the mean of the attribute.
To calculate the probability of differentiation, each value on the scale is initially considered a separate variable. Each possible response is coded as a numerical value. The absence of a dot is coded as zero, the smallest dot as one, the medium dot as two, and the largest dot as three. The scores of the six scale values are summed. Then the score of each scale value is divided by the sum to arrive at the proportion of the category members perceived to be described by each level of the attribute (pi). Probability of differentiation for each attribute is calculated according the following formula:
Probability of Differentiation = 1 - S i=1,6 pi2
The scores of the four different attributes are averaged to create a single score for each society. The reliabilities for these items are .962 for those evaluating the US, and .934 for those evaluating Taiwan.
To calculate perceived variability, each position on the scale was assigned a scale value of one to six. For example, in reference to the scale for adventurousness, the "very adventurous" position was assigned a scale value of six, "moderately adventurous" was assigned a scale value of five, and so forth down to "very unadventurous," which was assigned a scale value of one. The means (M) of the perceived distributions were calculated for each scale through the following formula:
M= S i=1,6 pi Xi
X is the scale value and p is the subjectively perceived proportion of category members described by level i of the attribute. Perceived variation for each attribute was calculated according to the following formula:
Perceived Variability = S i=1,6 pi(X i - M)2
The scores of the four different attributes were averaged into single score for each society.
The original scales Linville and her colleagues developed asked participants to write the proportion of the members of the group who would fall into each of the six values rather than using the stick-on dots to indicate these proportions. These scales have been used in a variety of studies (e.g., Linville, Fischer, and Salovey, 1989; Judd and Park, 1990; Lee and Ottati, 1995; Thompson, Kohles, Otsuki, and Kent, 1997). Although these scales seem to capture perceived dispersion, they are awkward and time consuming to complete. Park and Judd's (1990) modification of this measure makes the procedure easier for the participants and therefore more appropriate for inclusion within a lengthy questionnaire, while providing results that are consistent with those of the original measure. The reliabilities of these items for those evaluating the US were .723 and .841 for those evaluating Taiwan.
Manipulation checks and priming summary evaluation. At the end of the questionnaire, participants were asked "when they were first aware of when the film was made?" Participants circled one of four options that ranged from 1 "before seeing the clip" to 4 "never."
After evaluating the film, participants were asked to complete were two seven-point scales evaluating the "summary you read before you saw the film." "Based on what you've seen of the film so far" they were asked, "how accurate do you think the summary is? That is, do you think the summary is correct in its evaluation of the film?" The ends of the scale were marked "not at all accurate"(1) and "very accurate" (7). Then, participants were asked to evaluate how relevant they thought the summary was. "That is, [did they] think the summary discusses the most important things about the film, or does it discuss things that were relatively unimportant?" The ends of the scale were marked "not at all relevant" (1) and "very relevant" (7).
I predicted that the perceived representativeness of the characters would influence the degree to which the characters' attributes were applied to the audience's understandings of the film society. The more representative a character is seen to be, the more likely the audience should be to construct their perceptions of the film society using the film characters as a model. Perceived character representativeness should be associated with perceptions of the film society that concord with perceptions of the characters. I also predicted that the participants who read the film summary that emphasized the films' nationality will see the portrayed society as more homogeneous and the film characters as more representative than those who read the control version. Furthermore, I predicted that the perceived variability of the society would be negatively associated with the degree to which the traits of the characters are applied to their real-world understandings of the portrayed society. Since the priming manipulation has the potential to affect measures of character representation in ways that may affect the investigation of other hypotheses, I report these results first. Then, I discuss the impact of character representativeness on the audiences' application of the character's attributes to perceptions of the characters' real-world societies.
I predicted that the participants who read the priming summary would describe the film society as less variable than would those who read the control summary. Furthermore, participants who read the priming summary should see the characters as more representative than those who read the control summary.
Manipulation check. Analysis of the manipulation check in regards to the US film indicates that those who read the control summary felt they were aware of where the film was made later than those in the control condition, t(57)= -2.44, p=.02. The manipulation was successful, if not strong.
The manipulation check in relation to the Taiwanese film indicates that there is no significant difference between the primed and non-primed viewers in terms of when they found out where the film was made, t(56)=-.63, p=.53. This suggests that the summaries presented before the film clip may have been unsuccessful in activating the category structure of society membership among the participants. This result implies that the manipulation may not affect the participants' perceptions and suggests that any lack of results may be due to a failure of the manipulation.
I compared participants' evaluations of the relevance and accuracy of the priming and control summaries. There were no differences in the evaluations of the accuracy or relevance of the priming and control summaries in regard to either film. In both cases, the control and priming summaries are viewed as equally accurate and relevant.
Manipulation effects. I predicted that those in the primed condition would have greater homogeneity ratings of the films' societies than those in the control condition. Among those who evaluated the US, the overall mean of the range measure is 12.56(SD=1.54) on a scale that ranges from 0 to 15.2. The range is designed to capture the audiences' sense of the breadth of variation within the population on a set of attributes. The second measure of variability is the probability of differentiation, which represents the average probability across attributes that two randomly chosen group members will vary in terms of the attribute. The overall mean for this measure was .78(SD=.11) in relation to the US. Perceived variation is the average, across attributes, of the extent to which category members are perceived as widely dispersed about the mean of an attribute. It is basically the average of the variation of the scales constructed by the participants. The overall mean of the perceived variation measure is 2.37(SD=.49).
Although the directions of the difference in the means between the control and priming groups is in the predicted direction for all three measures of variability, the difference does not reach statistical significance for any of the measures, range, t(54)= -.22, p=.83; probability of differentiation, t(57)= -.46, p=.65; perceived variation, t(57)= -1.91, p=.06. The means within the primed and control groups for each of these measures are reported in Table 1.
I sought to determine if perceptions of the variability of the society are associated with perceptions of the representativeness of the characters. There were no direct priming effects on the perceived representativeness of either character, Beth, t(57)= -.333, p=.74; Nick, t(57)=.450, p=.65 or of the characters evaluated together, t(57)=.054, p=.96. Character representativeness is not significantly associated with perceptions of the variability of the host society.
In relation to the Taiwanese film, the overall mean the range index is 10.98(SD=2.36). The mean of the probability of differentiation measure is .774(SD= .132), and the mean of perceived variability is 2.14(SD= .59). However, the results of t-tests checking for priming effects of these measures are not significant. There were no significant differences in any of the three variability measures. The means of the two different groups on these measures are reported in Table 2. Although the difference in the range measure was in the predicted direction, measures of perceived variability and probability of differentiation were greater among the control group than the primed group. This is contrary to prediction.
The hypothesis that priming social category memberships would affect perceptions of the representativeness of the society is not supported, nor is the hypothesis that perceived variation would be associated with perceived representativeness.
Potential moderators. I sought to determine whether there were any other factors that might be moderating or masking the impact of the priming manipulation on perceptions of society variability. I tested for interaction effects between the priming manipulation and four types of control variables, 1) the participant's previous exposure to the films' cast, 2) their frequency of film viewing, 3) their reaction to the film and its characters, and 4) the demographics of the participant in relation to those of the characters.
Cast recognition and frequency of film viewing were found to be associated with perceptions of character representativeness in pretests, and so they are investigated as potential interactions here. Reactions to the film itself, such as comprehension and interest, could also have shaped the impact of the prime by influencing the audiences' degree of engagement with the film. Demographic factors are also considered because points of similarity other than society membership, including gender, age, race or ethnicity, could impact the audiences' responses to the characters. Similarity between the character and the participant in terms of these factors could impact components of interpretation such as attention and identification that may shape representativeness perceptions. The prime, therefore, may affect those who have some other connection to the characters differently than those that do not. Furthermore, if the characters are evaluated in terms of these category memberships, it could obscure any effects of the salience manipulation.
These factors were investigated through ANOVA analyses predicting the perceived representativeness of the characters that incorporated interaction terms along with the main effect variables. The representativeness of neither character in either film is predicted by the interaction of the prime and participants' interest in the film, their subjective sense of understanding of the film, their perception of the film's genre, or their sense of identification with the characters. There are no significant interaction effects between the priming manipulation and the age, race or gender of the participant, their frequency of general movie watching or the frequency of watching foreign movies or perceptions of the strength or goodness of the characters.
Impacts of Representativeness on Real-world Perceptions There were two characters evaluated in the US film, a young woman named Beth and a young man named Nick. On an index that ranges from one to seven, the overall mean of the representativeness evaluations for Beth is 4.07(SD=1.19), whereas the average for Nick is 4.10(SD=1.12).
According the pretests, the attributes that viewers associated with Beth are polite, unambitious, self-controlled, unconfident, practical, quiet, serious, and shy. Using OLS regression with listwise deletion, I sought to determine whether the perceived representativeness of this character would be significantly associated with the degree to which these eight attributes were applied to the audiences' evaluations of young US women. There were no significant associations. There were not any significant associations between representativeness and the degree to which the attributes were used to describe the "average person" in the US. I tried controlling for how much the participant reported liking the character, their evaluations of the "goodness" of the character, and their level of identification with the character. These analyses did not reveal any significant associations.
The attributes that were associated with Nick at the .01 level were impolite, modern, unambitious, impulsive, confident, idealistic, and individualistic. I sought to see whether the perceived representativeness of this character would be significantly associated with the degree to which these seven characteristics were applied to the audiences' evaluations of young US men. The only significant association is that the representativeness of Nick negatively predicted the degree to which young US men were seen as individualistic, B= -.397(.20), p=05. This effect is in the direction opposite that predicted. I also sought associations between the perceived representativeness of this character and the "average person" in the US. There were no significant associations. Controlling for liking the character, evaluations of the goodness of the character, and identification with the character did not reveal any significant associations between representativeness and evaluations of the character's subgroups.
Two characters in the Taiwanese film segment were evaluated. Each was a young woman. One was called Jia-Chien and the other Jia-Ning. The mean of the representativeness index is 4.88(SD=1.03) for Jia-Chien and 4.22(SD=1.25) for Jia-Ning. According to the pretests, the attributes associated with Jia-Chien are impolite, modern, ambitious, confident, individualistic, and serious. Using OLS regression, I checked to see whether the representativeness of this character was associated with greater perceptions that people within her subgroups would carry these attributes. None of the associations between representativeness and evaluations of the character's subgroups are significant.
The traits associated with Jia-Ning are polite, ambitious, talkative, and happy-go-lucky, and friendly. I checked to see if the representativeness of this character predicted evaluations of her subgroup within the film setting society. There was only one significant result. The representativeness of the character of Jia-Ning is positively associated with the evaluations of the friendliness of young Taiwanese women, B=.333(SD=.15), p=.03. Friendliness is an attribute that was strongly associated with the character in the pretests, but which was not strongly associated with perceptions of her subgroup of Taiwanese society. The attribute is not stereotypical. Furthermore, this character was most liked, mean=5.39(SD=1.35) and the most identified with, mean=4.05(SD=1.71) of the four characters studied. The fact that this attribute, in relation to this character, represents the only significant result suggests some interesting avenues for investigating the circumstances under which media may have a positive, counter-stereotypical effect on audiences' perceptions of other countries. However, given the number of tests that this analysis represents, the confidence one can have in it is limited.
I also sought to determine whether the perceived representativeness of either of these characters was associated with evaluations of the "average person" in Taiwan. I controlled for liking the characters, the perceived goodness of the characters, and identification with the characters. These analyses did not reveal any further associations between representativeness and evaluations of the characters' subgroups. This analysis provides little support for the prediction that character representativeness would be associated with the degree to which the characters' attributes are applied to their real-world groups.
The results hint at situations in which media representations may shape audiences' real-world perceptions. Specifically, one character impacted perceptions of her society when the society was socially distant from the audience, the audience connected with and liked the character, and the character's attributes were different from the prevailing society stereotype. Future research may benefit from investigations into the way cross-cultural media identification is established. These connections may be a means through which media representations could produce positive change on audiences' perceptions of their societies.
However, all things considered, the results of this study are sparse. Its most valuable lessons come as warnings and guides of future research in this area. There are several possible reasons why this study failed to support the hypotheses it examined. The audiences' perceptions of the film societies may have been too well established to be affected by these brief film segments. When one examines the failure to find any differences across the experimental and control groups, the most obvious explanation is that the priming materials did not successfully activate the category structure of society membership. When one considers the failure to find an association between attribution and the application of the characters' attributes to the character's subgroups, it may be that the traits revealed by the characters' behavior were not captured in the measures of the characters' effect on the participants' real-world understandings. For example, if a participant felt that a character's action revealed his or her insensitivity, but the character's impact on the audiences' perceptions of the real-world group's insensitivity was not measured, the study would have failed to capture an aspect of the representation's effect. It is also possible, of course, the hypotheses simply incorrect. The literature from which the study questions were adapted was based on interpersonal contexts. The way individuals process information about the media may be different from the way they process interpersonal information in ways that I did not anticipate.
Austin, B.A. (1989). Immediate Seating: A Look at Movie Audiences. Belmont, CA: Wadsworth.
Beckett, N.E., and Park, B. (1995). "Use of category versus individuating information: making base rates salient." Personality and Social Psychology Bulletin, 21, 21-31.
Citipitioglu, L., and Elasmar, M.G. (1996). "The effects of imported American television on the attitudes of Turkish youth towards to United States." Paper presented at the International Communication Association, Chicago, IL.
Cohen, C.E. (1981) "Person categories and social perception. Testing some boundaries of the processing effects of prior knowledge." Journal of Personality and Social Psychology, 40, 441-452.
Custen, G.F. (1980). "Film talk: Viewers' responses to film as a socially situated event." Unpublished doctoral dissertation, Annenberg School for Communication, University of Pennsylvania.
Darley, J.M., and Gross, P.H. (1983). "A hypothesis-confirming bias in labeling effects." Journal of Personality and Social Psychology, 44, 20-33.
Devine, P.G. (1989). "Stereotypes and prejudice: Their automatic and controlled components." Journal of Personality and Social Psychology, 56, 5-18.
Dovidio, J.F., Evans, N., and Tyler, R.B. (1986). "Racial stereotypes: The contents of their cognitive representations." Journal of Experimental Social Psychology, 22, 22-37.
Duncan, B.L. (1976). "Differential social perception and attribution of intergroup violence: Testing the lower limits of stereotyping of Blacks." Journal of Personality and Social Psychology, 34, 590-598.
Ford, T.E., Stangor, C., and Duan, C. (1994). "Influence of social category accessibility and category-associated traid accessibility on judgements of individuals." Social Cognition, 12, 149-168.
Gaertner, S.L., and McLaughlin, J.P. (1983). "Racial stereotypes: Associations and ascriptions of positive and negative characteristics." Social Psychology Quarterly, 46, 23-30.
Gilbert, D.T., and Hixon, J.G. (1991). "The trouble of thinking: Activation and application of stereotypic beliefs." Journal of Personality and Social Psychology, 60, 509-517.
Haslam, S.A., Oakes, P.J., Turner, J.C., and McGarty, C. (1995). "Social categorization and group homogeneity: Changes in the perceived applicability of stereotype content as a function of comparative context and trait favorableness." British Journal of Social Psychology, 34, 139-160.
Heung, M. (1997). "The family romance of orientalism: From Madam Butterfly to Indochine." In M. Bernstein and G. Studlar (eds.) Visions of the East: Orientalism in Film, p. 158-183. New Brunswick, N.J.: Rutgers University Press.
Hilton, J.L., and von Hipple, W. (1996). "Stereotypes." Annual Review of Psychology, 47, 237-271.
Johnston, L., and Hewstone, M. (1992). "Cognitive models of stereotype change." Journal of Experimental Social Psychology, 28, 360-386.
Judd, C.M., Park, B., Ryan, C.S., Brauer, M., Kraus, S. (1995). "Stereotypes and ethnocentrism: Diverging interethnic perceptions of African American and White American youth." Journal of Personality and Social Psychology, 65, 460-481.
Kippax, S. and Brigden, D. (1977). "Australian stereotyping--a comparison." Australian Journal of Psychology, 2, 89-96.
Kunda, Z., and Oleson, K.C. (1995). "Maintaining stereotypes in the face of disconfirmation: Constructing grounds for subtyping deviants." Journal of Personality and Social Psychology, 68, 565-579.
Kunda, Z., and Oleson, K.C. (1997). "When exceptions prove the rule: How extremity of deviance determines the impact of deviant examples on the stereotypes." Journal of Personality and Social Psychology, 72, 965-979.
Ladonde, R.N., and Gardner, R.C. (1989). "An intergroup perspective on stereotype organization and processing." British Journal of Social Psychology, 28, 289-303.
Lee, Y., and Ottati, V. (1995). "Perceived in-group homogeneity as a function of group membership salience and stereotype threat." Personality and Social Psychology Bulletin, 6, 610-619.
Liebes, T., and Katz, E. (1990). Exporting Meaning: Cross-cultural Readings of Dallas. New York: Oxford University Press.
Linville, P.W., Salovey, P., and Fischer, G.W. (1986). Stereotyping and perceived distributions of social characteristics: An application to ingroup-outgroup perception. in J. Dovidio and S.L. Gaertner (eds.), Prejudice, Discrimination, and Racism. (pp. 165-208). New York: Academic Press.
Linville, P.W., Fischer, G.W., and Salovey, P. (1989). "Perceived distribution of the characteristics of in-group and out-group members: Empirical evidence and a computer simulation." Journal of Personality and Social Psychology, 57, 165-188.
Livingstone, S.M. (1990). Making Sense of Television: The Psychology of Audience Interpretation. Oxford: Pergamon.
Morley, D. (1980). The Nationwide Audience: Structure and Decoding. British Film Institute. Television Monograph no. 11. London: BFI.
Nadel, A. (1997). "A whole new (Disney) world order: Aladdin, atomic power, and the Muslim Middle East." In M. Bernstein and G. Studlar (eds.), Visions of the East: Orientalism in Film, p. 184-203. New Brunswick, NJ: Rutgers University Press.
O'Sullivan, C.S., and Durso, F.T. (1984). "Effect of schema-incongruent information on memory for stereotypical attributes." Journal of Personality and Social Psychology, 47, 55-70.
Park, B., and Hastie, R. (1987). "Perception of variability in category development: Instance- versus abstraction-based stereotypes." Journal of Language and Social Psychology, 53, 621-635.
Park, B., and Judd, C.M. (1990). "Measures and models of perceived group variability." Journal of Personality and Social Psychology, 59, 173-191.
Pingree, S. and Hawkins, R. (1981). "US programs on Australian television: The cultivation effect." Journal of Communication, 31 (1), 97-105.
Rehder, B., and Hastie, R. (1996). "The moderating influence of variability on belief revision." Psychonomic Bulletin and Review, 3, 499-503.
Rothbart, M., Evans, M., and Fulero, S. (1979). "Recall for confirming events: Memory processes and the maintenance of social stereotypes." Journal of Experimental Social Psychology, 15, 343-355.
Ryan, C.S., Judd, C.M., and Park, B. (1996). "Effects of racial stereotypes on judgments of individuals: The moderating role of perceived group variability." Journal of Experimental Social Psychology, 32, 71-103.
Sagar, H.A., and Schofield, J.W. (1980). "Racial and behavioral cues in Black and White children's perceptions of ambiguously aggressive acts." Journal of Personality and Social Psychology, 39, 590-598.
Said, E.W. (1978). Orientalism. New York: Random House.
Saito, S. (1996). "The image of the partner: Television's contribution to Japanese perceptions of America." Unpublished doctoral dissertation, Annenberg School for Communication, University of Pennsylvania.
Shohat, E., and Stam, R. (1994). Unthinking Eurocentrism: Multiculturalism and the Media. London: Routledge.
Shome, R. (1996). "Race and popular cinema: The rhetorical strategies of Whiteness in City of Joy." Communication Quarterly, 44, 502-518.
Springer, C. (1991). "Comprehension and crisis: Reporter films and the Third World." In L.D. Friedman (ed.), Unspeakable Images: Ethnicity and the American Cinema, p. 167-189. Urbana, IL: University of Illinois Press.
Tan, A.S., Li, S., and Simpson, C. (1986). "American TV and social stereotypes of Americans in Taiwan and Mexico." Journalism Quarterly, 63, 809-814.
Tan, A.S. and Suarchavarat, K. (1988). "American TV and social stereotypes of Americans in Thailand." Journalism Quarterly, 65, 648-654.
Thompson, S.C., Kohles, J.C., Otsuki, T.A., and Kent, D.R. (1997). "Perceptions of attitudinal similarity in ethnic groups in the US: Ingroup and outgroup homogeneity effects." European Journal of Social Psychology, 27, 209-220.
Torgovnick, M. (1990). Gone Primitive: Savage Intellects, Modern Lives. Chicago: University of Chicago Press.
Weber, R. and Crocker, J. (1983). "Cognitive processes in the revision of stereotypic beliefs." Journal of Personality and Social Psychology, 45, 961-977.
Weimann, G. (1984). "Images of life in America: The impact of American TV in Israel." International Journal of Intercultural Relations, 8, 185-197.
Wilder, D.A., Simon, A.F., and Faith, M. (1996). "Enhancing the impact of counterstereotypic information: Dispositional attributions for deviance." Journal of Personality and Social Psychology, 71, 276-287.
Willnat, L., He, Z., and Hao, X. (1996). "The effects of foreign media exposure on cognitive and affective perceptions of Americans in Hong Kong, China, and Singapore." Paper presented at the International Communication Association, Chicago, IL.